This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.
An optimized fat-water multi-parametric imaging which combines conventional sequences with some of the advanced MRI techniques could enable a comprehensive examination of the liver, providing information on the presence of fat, iron, and fibrosis as well as hepatocellular carcinoma (HCC) and portal hypertension and could represent the future of abdominal imaging, could possibly replace liver biopsy, or at least be a mechanism for monitoring therapeutic progression of chronic liver disease. The migration to clinical 3T MR imaging can provide higher signal-to-noise ratios and should in theory improve image quality. T2-contrast images have their maximal signal intensity in fluid-filled regions, which is important in contrasting tissues in clinical applications. Fatty liver disease is the most common cause of chronic liver disease in both children and adults; but fat is said to produce high signal intensity on routine single short fast spin-echo (SSFSE) sequences which may obscure pathological causes of the high signal emanating from structures that are surrounded by fat, include organs such as bowel, pancreas, and kidneys. It has been shown that many of these pathological conditions show up most readily under T2-contrast images, and as such T2-weighted images are used in conjunction with proton-density-weighted or T1-weighted images as anatomical references in many advanced MRI studies. T2-weighting follows directly from the spin density weighting sequence and the signal is usually the second echo which is produced by a second 180-degree pulse in a given TR spin echo pulse sequence. The influence of T2 is minimized in a short TE and as such T2-contrasted images cannot be obtained at short echo time (TE), but at intermediate echo times at which the differences in magnetization between tissues can be maximized. T2-weighted images can be generated using spin-echo based pulse sequences which allow true spin-spin relaxation independent of the field inhomogeneity. Optimal echo spacing is way to provide better signal-to-noise ratio (SNR) in shorter scan time. Echo space is the distance in time between the echoes in multiple echo sequences and shorter echo space produces compact sequence timing, more slices and better resolution and possibly reduced artifacts. Given hardware manufacturing limitations in multi-echo imaging measurements and that the time required for sampling each echo is typically on the order of milliseconds; there is often a lower limit on the first echo and as such the spacing for the second and subsequent echoes.
Inversion Recovery (IR):
While inversion-recovery (IR) techniques have favorable characteristics in providing more robust uniformity and are generally less sensitive to susceptibility affects; a potential drawback to its most common implementation is the use of wide frequency bandwidth which may include both fat and water spins, and which has resulted in water signal not being fully recovered during data acquisition, and overall water signal-to-noise ratio (SNR) being diminished. The image formed from the first TE is usually proton density weighted, and as TE is increased, more T2 contrast is achieved, but at the expense of reduced transverse magnetization signal. Echo spacing has undetermined impact on the direct calibration of T2 relaxation values. The objective is to investigate the effect of echo spacing on the quantification of T2 relaxation values using 3.0 Tesla machine as a precursor for further investigation of its application in abdominal imaging for direct fat quantification in fatty liver diseases.
There are many causes of advanced liver failure or chronic liver disease including nonalcoholic fatty liver disease (NAFLD), excessive use of alcohol, viral infections, primary sclerosing cholangitis, primary hemochromatosis, and autoimmune disease. Chronic liver diseases can lead to hepatic fibrosis, cirrhosis, end-stage liver disease, portal hypertension, hepatocellular carcinoma (HCC) and these can be the primary cause of many morbidity, mortality, as well as constituting as major drain on national health care resources. In the United States, two primary causes of chronic liver disease are hepatitis C virus (HCV) infection and nonalcoholic fatty liver disease (NAFLD).
It is estimated that hepatitis C virus (HCV) infections account for about 40% of all the chronic liver diseases, and is the most frequent cause for liver transplantation, as well as causing an estimated 10,000 deaths per year [1,2]. The incidence of nonalcoholic fatty liver disease (NAFLD) is increasing with the growing worldwide epidemic of obesity. In United States the prevalence of nonalcoholic fatty liver disease has increased dramatically, which reflects the obesity epidemic that afflicts an estimated 100 million people including 10% of children. With these increasing epidemics of obesity and diabetes, nonalcoholic fatty liver disease may overtake chronic HCV infection as the leading cause of chronic liver failure within the next decade. Nonalcoholic fatty liver disease (NAFLD) encompasses a spectrum of conditions associated with lipid deposition in hepatocytes, and ranges from steatosis (simple fatty liver), to nonalcoholic steatohepatitis (NASH), and to advanced fibrosis and cirrhosis. Simple fatty liver is believed to be a benign condition, but NASH can progress to fibrosis and may lead to end-stage liver disease. Patients with nonalcoholic steatohepatitis have higher levels of aspartate aminotransferase and ferritin, more hepatocyte ballooning, and more fibrosis in histologic sections [27-29]. The pathogenesis of HCV, NAFLD and NASH varies in many ways, but may ultimately lead to accumulation of both fats and irons in the hepatocytes. The accumulation of these elements in the liver, in addition to inflammation, may lead to irreversible damage to liver cells; thus contributing to fatty liver disease being the most common cause of chronic liver disease in the United States.
Identification and quantification of chronic liver diseases have been extensively studied, using blood tests as marker for liver function, liver biopsies and most of the non-invasive diagnostic imaging, including Ultra sound, computed tomography (CT), conventional and advanced MRI. Hepatic fibrogenesis is a complex dynamic process, which is mediated by necroinflammation and activation of stellate cells with abnormal collagen deposition . Liver function test has not shown to correlate well with the degree of fibrosis. Serologic markers of hepatic fibrosis, such as aspartate aminotransferase (AST)/ALT ratio, platelet count, and prothrombin index, are said to have variable accuracies . Another previous study showed that up to 40% of patients with advanced fibrosis have persistently normal alanine aminotransferase (ALT) levels .
Liver biopsy is said to be the gold standard in the diagnosis of chronic liver diseases. Liver biopsy is useful in determining the presence of steatosis, steatohepatitis, assessing the degree of fibrosis and can also be helpful in ruling out other alternative diagnosis. Although liver biopsy is a relatively safe procedure when performed by experienced clinicians; it has poor patient acceptance, is not risk free, and the results are inherently subjective. The limitation of liver biopsies include the heterogeneity of disease distribution throughout the liver parenchyma, which result in substantial sampling variability, diagnostic and staging inaccuracies related to the random sampling of limited portions of the entire volume of the liver [9, 10]. A reliable and reproducible noninvasive marker of hepatic fibrosis is strongly needed, a tool that would reduce biopsy-related risks and costs, and which could be useful for guiding antiviral treatment and monitoring treatment efficacy, and can be used for clinical evaluation of new types of antiviral and antifibrotic drugs .
A liver ultrasound examination is useful for confirming steatosis, it has not been successful in ruling out steatohepatitis or fibrosis, and its sensitivity drops sharply when the degree of steatosis decreases below 30%. Due to improved CT image quality since the introduction of multi-detector, CT utilization has substantially increased for the screening of liver diseases . Unenhanced CT is considered the best CT method for estimation of liver fat; and it involves simple measurement of liver attenuation in Hounsfield units [19-22]. The measurement of attenuation is based on the physical characteristic of x-ray penetration of tissue. But with ionizing radiation, CT cannot be used on a prolonged basis to monitor disease prognosis and the response to therapy. Conventional MRI is biologically based may actually be more sensitive than CT, and without the effects of ionizing radiations, has become an increasingly important imaging technique for the investigation of patients with chronic liver disease. Recent advances in MRI have led to a growing interest in optimizing the MRI methods for assessment of chronic liver diseases. These MRI methods include but are not limited to versions of T1-weighted, T2-weighted, PD-weighted, diffusion-weighted imaging (DWI), perfusion-weighted MRI, MR elastography (MRE), and MR spectroscopy (MRS). However, many of the MRI sequences involve complex acquisition and processing, some desiring learning curve and in some cases the sequence may be unavailable for clinical use. Even though that most of MRI findings are also be subjective, and subject to interobserver variability and with varying degrees in sensitivity and specificity; it provides ample opportunity for vast improvement and wide acceptability.
The recognition of nonalcoholic steatohepatitis as a distinct diagnosis has increased interest in noninvasive quantification of both liver fat and iron deposit with MRI imaging, since such measurements might serve as markers of the disease severity and therapeutic response. Many of the diagnoses and quantifications of hepatic steatosis with MRI sequences exploit the fact that fat resonates more slowly than water (210 Hz at 1.5 T). Among many MRI sequences, MRS is considered to be the reference standard method for measuring hepatic fat content, with many studies showing strong correlations between MRS and known grades of steatosis. MRS uses single-voxel spectroscopy approaches which interrogates a large voxel (typically 2.0 x 2.0 x 2.0 cm, or 8.0 cm3), but this cannot provide accurate volumetric evaluation of the fat content of the liver. It is well known that hepatic steatosis is often heterogeneous, and a single-voxel spectroscopy cannot provide a comprehensive evaluation of hepatic steatosis. Other studies have suggested that in-phase and out-of-phase imaging can be used to rapidly estimate the liver fat content in patients with nonalcoholic fatty liver disease (NAFLD). It has also been shown that cirrhosis is associated with increased liver iron deposit , which being a paramagnetic substance results in localized field inhomogeneity. Any such field inhomogeneity may cause T2* effects and paradoxical signal intensity loss on in-phase versus out-of-phase MR images because of the longer echo time of the in-phase images . Other MRI methods that have been developed for the quantification of hepatic iron overload are based on both T2*- and T2-weighted imaging that was initially developed by Gandon et al; it uses a combination of 2D gradient-echo images acquired with proton density (PD) weighting, T1 weighting, and escalating T2* weighting. Wood, et al. also has similar approaches of iron quantification that focused on direct measurement of T2* or R2* mapping using a 3D multi-echo gradient-echo acquisition . It has also been shown that fat demonstrates high signal intensity on T2-weighted fast spin-echo and single-shot fast spin-echo (SSFSE) MR images. The result being that fat quantification, by both fast spin-echo and single-shot fast spin-echo sequences could be more sensitive for detection of the presence of fat than are conventional spin-echo T2-weighted sequences . T2-weighted fast spin-echo MR imaging without fat suppression and with fat suppression appears to provide one of the purest evaluation of fatty infiltration, without the confounding effects of T2*, fibrosis, and other factors that may limit assessment using in-phase and out-of-phase MR imaging. T2W images are said to have their maximal signal intensity in fluid-filled regions, which is important in contrasting tissues in clinical application. It has been shown that many pathological conditions show up most readily under T2W images, and as such T2W images are used in conjunction with proton density (PD) or T1 weighted images as anatomical reference in many advanced MRI studies. Again with fat producing high signal intensity on routine SSFSE that may obscure pathological causes of high signal that emanate from abdominal structures that are surrounded by fat, including organs such as bowel, pancreas, and kidneys. Many of these diseases include but not limited to inflammation, infection, and abnormal fluid collections. Inversion-recovery (IR) techniques are said to have favorable characteristics in providing more robust uniformity and are generally less sensitive to susceptibility affects from sources such as bowel gas. Three MR imaging techniques that permit the breakdown of the net MR signal into fats and water signal components are in clinical use today for detection and quantification fats by utilizing the difference in precessional frequency between water and fats. A retrospective study by Aliya Qayyum, et al. suggested that liver fat may be more accurately quantified with fat-saturated FSE imaging than with out-of-phase gradient-echo MR imaging, especially in patients with cirrhosis.
Basic Physics of MRI & Spin Echo
The main magnetic field of an MR scanner comes from a large electric current flowing through wires that are formed into a loop in the magnet of the imaging system. These wires are immersed in liquid helium (at superconducting temperatures) so that very large currents can be used to produce the strong magnetic field and the liquid helium levels in the magnet is refilled at regular intervals. Power supply can be used to inject electric current into the coils of wire (ramped), and this initial power supply will then be removed. The magnetic field of the imaging system can retain this electric current for many years with no additional electric current and with only minimal loss in electric current and minimal decrease in magnetic field strength. It is necessary to have a source of hydrogen protons in order to generate MR signal and the signal from a tissue is determined in large part by its hydrogen concentration.
This is the concentration of resonating hydrogen atoms in a given molecular structure. The MR signal obtained from a tissue is determined in large part by its hydrogen concentration. Hydrogen proton is positively charged and spins about its axis and as such acts like a tiny magnet. The hydrogen protons in human body and similar biological tissues, thus act like many tiny magnets. Hydrogen proton exists in different chemical form in the body, mainly in water and fat. When a patient is placed in an external magnetic field, or similarly if any biological phantom is placed in the MRI scanner, the hydrogen protons begin to align in given directions dictated by the laws of quantum physics. It turns out that in the case of the hydrogen nuclei that two discrete energy levels were created; a higher energy level with magnetic moments opposing the external magnetic field, and a lower energy level in which the nuclei are aligned in the direction of external magnetic field. Therefore a net magnetization is produced following the application of an external magnetic field that will cause a small majority of spins to align in the direction of the applied field. The population difference between high and low energy levels which is the measure of sensitivity, can be altered by increasing the field strength, hence the desire for a strong magnetic field, which for modern clinical scanners can be 1.5 or 3.0 Tesla and many research scanners can range from 3.0 - 11.7T or even more. In order to detect this signal which is a reflection of energy difference, a second magnetic field must be applied perpendicular to the external magnetic field (z direction usually referred as B0) at the same resonant frequency (RF) as resonance is said to aid in an efficient transfer of energy.
It is import to note that spins aligned with main magnetic field do not contribute any signal toward the resulting image. To generate any image, the desired spins must be flipped into the transverse plane. The basic sequence of spin echo consists of a 90-degree RF pulse excitation, spatial encoding, a 180-degree refocusing pulse, and signal readout. Note that only spins, precessing at a frequency equivalent to that of the 90-degree RF pulse will be flipped into the transverse plane. In a fast spin echo sequence, there are multiple phase encoding steps, each applied for each single echo during the long TR time. All data collected are used in generating an image and the effective TE is used to specify the contrast. Effective Echo Time (ETE) is the relative echo time seen on the final image generated from the echo train sequence. The strength and duration of the RF pulse can be controlled to rotate the net magnetization to any angle and if the RF pulse rotates the net magnetization 90Â° into the transverse plane, it is termed a 90-degree RF pulse. If the RF pulse rotates the net magnetization 180Â° into the -z direction, that is termed a 180-degree RF pulse. Both 90-degree and 180-degree RF pulses are important when discussing the techniques associated with both spin echo (SE) and gradient-recalled-echo (GRE) imaging. The net magnetization that is aligned with the main magnetic field is usually called longitudinal magnetization; application of 90Â° RF pulse will rotate this longitudinal magnetization into the transverse plane, and this is called transverse magnetization (B1). Using appropriate RF coils to transmit B1 signal, this acts to tip the spins out of alignment with external magnetic field (B0) and towards the direction of the coil (i.e. from longitudinal plane, towards the transverse plane). The spin echo can be used to recover all dephasing signal except for spin-spin interactions. After a 90Â° RF pulse, protons that were in phase in the transverse plane will begin to dephase. And after a certain amount of time, if a 180Â° RF pulse is applied, the spins will rotate over to the opposite axis; and rather than the spins continuing to dephase, the spins will begin to rephase. The spins will come back together to form a maximum signal measured with the receiver coil; after which spins begin to dephase again until another application of 180Â° RF pulse. The process can continue until no signal can be recovered.
A signal peak is formed at the echo time (TE), when the transverse magnetization is completely in phase and the measured MR-signal is at its maximum. When the transverse magnetization begins to dephase, the measured MR-signal begins to decrease until the magnetization is completely dephased or out of phase, at which time the measured signal is zero. The primary factors that provide image contrast in a typical spin-echo sequence are proton density, longitudinal relaxation (T1), and transverse relaxation (T2). Although proton density which constitutes brightness modulation is always present in all the images, further contrast can be manipulated by changing the timing parameters utilized in the sequence. Through modification of repetition time (TR), echo time (TE) and various gradient moments, image contrast can be altered to emphasize a particular relaxation time or the proton density. At long TR and minimum TE values, tissue contrast is solely from proton density as the difference in intensity that arises from relaxation times vanishes. Similarly images formed from short TR and long TE have very low signal-to-noise ratio (SNR) and minimum contrast and are generally not recommended. Tissues with high T1 values also tend to exhibit high T2 values and as such elevated proton densities. T1 relaxation is the rate at which longitudinal magnetization grows back and is different for protons associated with different tissues; this is the fundamental source of contrast in T1-weighted images. T1 is a parameter that is characteristic of specific tissue and varies with the strength of the main magnetic field. Recall that the net magnetization is made up of contributions from many protons, which are all precessing. During and Immediately after RF pulse, the protons are precessing together ("in phase"). But shortly after, begin to dephase due to several effects. In the case, when dephasing is purely due to the effect of spin-spin interactions, the dephasing is called T2 decay or T2 relaxation. T2 is also a parameter that is characteristic of specific tissue and characterizes the rate of dephasing for the protons associated with that tissue.
T1 and T2 relaxation times are base on the thermal motion or random collision of atomic particles within molecules of living tissues. These random motions cause fluctuations in energy levels within the hydrogen nuclei when placed in an external magnetic field. One of the advantages of MRI is its excellent soft-tissue contrast which can be manipulated in an infinite ways; by increasingly altering the echo time (TE) and holding steady the repetition time (TR), the signal contrast can be altered or T2-weighted. If a long TE is used, the inherent differences in T2 times of tissues will become apparent; signal from tissues with long T2 (e.g. water) will take longer to decay and their signal will be greater (appear brighter in an image) than the signal from tissues with short T2 (e.g. fat). T2 is again defined as the time that it takes for the transverse magnetization to decay to 37% of its original value; and different tissues have different T2 values and they are said to dephase at different rates. If it is possible to create an image at a time when the transverse magnetization curves were widely separated, then there would be high contrast between the tissues in the image. In a similar manner TR governs T1-contrast such that, tissues with long TR (e.g. water) will take longer time to recover back to their equilibrium magnetization values, and such that a short TR interval will make the tissue with long T1 appear dark compared to another tissue with short T1 (e.g. fat). It is important to note that both T1 and T2 relaxation processes occur simultaneously and it takes only a few seconds for most of the transverse magnetization to dephase and for most of the longitudinal magnetization to grow back to its original strength and parallel to the main magnetic field. Signal intensity difference arising from relaxation time differences at long TR and short TE, and tissue contrast would be due to proton density difference between the tissues. Images formed from short TR and Long TE tends to have very low signal-to-noise ratio (SNR) and negligible contrast and should be of no clinical value.
Historically, clinical MR imaging is based on the identification of signals that are produced from hydrogen, which are contained in water and fat components of human tissues. An important technique in MRI is removing the high signal fat component from the image (fat suppression). There are many ways this can be achieved but each method relies on either the resonant frequency (chemical shift) or relaxation time differences between water and fat. In the chemical selective saturation method a preparatory pulse sequence is acquired which uses narrow bandwidth RF pulse to excite the fat peak alone. The excited fat is then deliberately dephased leaving only the water available for subsequent detection. Another common method is the Short TI Inversion Recovery (STIR). An inversion-recovery pulse sequence is useful for suppressing unwanted signals in MR images, while contrast weighting is being controlled through selection of TR and TE. Spectral presaturation with Inversion Recovery (SPIR), is a combination of the two previous methods, only the fat is excited and then inverted as in the STIR method. This method is different version of the one first developed by Dixon which involves acquiring images with fat and water in phase and out of phase and performing an image subtraction. Ideally two sets of spin-echo images are obtained; one with the use of the presaturation pulse (fat-saturated images) and another without its use (non-fat-saturated images). These two sets of images would be obtained with identical imaging parameters except for application of the presaturation pulse, so that the two images are directly comparable. On a pixel-by-pixel basis, the non-fat-saturated image will contain the signal intensity of the water plus fat protons, whereas the fat-saturated image will contain only the signal intensity of the water protons. If fat-saturated and non-fat-saturated images are acquired with otherwise identical imaging parameters, the signal intensity differences between the two images can be wholly attributed to the presence of fat. SPIR technique uses a 180 degree pulse to selectively invert and null fat signal from the imaging volume. SPIR is much less susceptible to homogeneity of the main magnetic field as it depends on the T1 recovery times rather than the precessional frequency, and relaxation times are not affected by small changes in homogeneity. SPIR technique is important in differentiating pathology (water based) and normal tissue (fat based). Water or fat saturation is most effectively achieved on high field systems, hence the need for 3T MRI system.
Abdominal Imaging at 3T
Magnetic resonance (MR) imaging at 3T has offered an improved signal-to-noise ratio (SNR) compared with that at 1.5 T. However the use of 3T MR imaging for abdominal evaluations, has lagged behind because of the difficulty of imaging a large volume during respiratory motions. The physics of high field strength also bring some disadvantages, such as increases in the specific absorption rate, in magnetic field inhomogeneity effects, and in susceptibility artifacts, as well as the need for radiofrequency (RF) modifications. Images acquired with 1.5T MR systems were of good diagnostic quality, but the desire to reduce the scan time by increasing the speed and improved image quality required the migration to higher field strength. The improved SNR cannot be viewed in isolation from other imaging parameters; theoretically SNR can also be improved by increasing the imaging time or by increasing the number of signals averaged. T1 and T2 are intrinsic properties but they change substantially with field strength, and there are substantial differences between T1- and T2-weighted pulse-sequences used for 3T and 1.5T. A misregistration artifact commonly occurs where the signals from fat and water are superposed; the shift is said to be proportional to the magnetic field strength . Increase chemical shift artifacts at 3T, may have adverse effects for diagnostic quality of the images. Even though, the trend toward higher field strengths has been driven by the desire to increase the speed for imaging and improved image quality (SNR); there are factors that tend to counteract these benefits. Two of most the important effects are the increases in tissue T1 values and in the radiofrequency energy (SAR) needed to magnetically excite the protons in tissue. The increased T1 values may require an increase in repetition time (TR) and an increase in imaging time as a result of decrease in signal intensity and tissue contrast. T2-weighted fast SE sequences uses multiple 180Â° refocusing pulses and as such SAR intensive; two sequences, hyperechoes and transition pseudo steady states (TRAPS) may be used to alter the refocusing pulses in a manner that reduces SAR without diminishing SNR. The T2-weighted sequences most commonly used in abdominal imaging are the fast SE sequence and its variants, the Half-Fourier rapid acquisition with relaxation enhancement (RARE) sequence. Parallel imaging may employ shorter echo train to reduce SAR effects and to reduce image blurring that are characteristic of T2W fast SE imaging . Many benefits of 3T imaging are lost in many abdominal imaging cases due to standing wave artifacts which may severely degrade the quality of T2-weighted images obtained at 3.0 T in patients who are obese or have ascites. At 3T MR imaging, the energy (SAR) deposited by radiofrequency waves may be clinically significant, estimated to be fourfold compared to that at 1.5T. Standing wave effects are even more pronounced in pregnant women because of their large girth and the presence of amniotic fluid. Fetal MR imaging at 3.0 T are not recommended because of increased concern for fetal safety at high field strengths.
Using echo spacing to optimize TE values in a Multi-spin echo Fat Study
Echo spacing (ES) determines the separation in milliseconds of the echoes in the echo train length (ETL). Echo train length (ETL) is the total number of echoes collected during one TR in a fast scan imaging sequence. Its contribution to imaging contrast is based on the number of echoes used to generate an image and the time when the echoes are collected during the sequence. Optimization of protocols at 3.0 T for both clinical and in vivo applications requires an appreciation of the changes in T1 and T2 relaxation times, which accompany any change in field strength. These relaxation times can affect both the spatial resolution and the signal-to-noise ratio; but more importantly can help determine the nature of contrast in MR images. These relaxation times can also determine the selection of timing parameters for the imaging pulse sequence, which may affect the total imaging times and image quality. T1 and T2 are known to change with magnetic field strength and magnetization transfer, which depends on water mobility and other tissue properties, such that it cannot be predicted accurately with a simple theoretical calculation. Therefore, accurate determination of proton relaxation times in a high-field MR imaging is essential in exploring the capabilities of high-field MR imaging system. A large number of techniques for measuring T1 and T2 relaxation times in tissues have been reported . These methods, however, often require very long imaging times, which makes them susceptible to imaging artifacts caused by patient motions. Thus, relaxation parameters in the human abdomen have been rarely studied in details. The purpose of this study is in-part to measure the effect of echo spacing on T2 relaxation times of known fractions of olive oil and water mixture phantom as part of other effort to develop a safety MR biomarker that could be translated to human abdominal disease studies. Many MRI studies have investigated the use of magnetic resonance relaxation times for tissue characterization. A number have been performed in vivo with clinical abdominal imagers. A lot of information has been collected; but the vast results so far have not established any role of quantitative tissue relaxation time measurements in the clinical abdominal disease diagnosis. And none to the best of our knowledge has stipulated the effects of echo spacing in relaxation time measurements. Generating T2-contrast images require collection of images using pulse sequences with long TR and intermediate echo time (TE). The amount of signal loss depends on the time between excitations (TR) and acquisitions or echo times. In order to obtain exclusive T2- weighted images, TR must be very long, such that longitudinal recovery is nearly complete and T1-contrast is minimized. Pure T2-weighted images can only be generated using spin-echo based pulse sequences which allow true spin-spin relaxation and are independent of the field inhomogeneity.
Methods and Materials
To study the effect of echo space on T2 relaxation times obtained in phantom experiments of mixed fractions of olive oil and water, using both standard and Spectral Saturation Inverse Recovery fat suppression techniques of Carr-Purcell-Meiboom-Gill (CPMG) imaging sequences.
MR Scanner and Phantom
The MR scanner is a 3.0-Tesla superconductive magnet (Achieva by Philips medical system, Netherlands). The phantom is made of nine cylindrical vials consist of 1 agarose gel, 1-pure-water, 1 pure olive oil, 6 mixed fractions of olive oil and water (10, 20, 30, 40, 50, & 60% of olive oil). Other fixed parameters include - Receiving Coil: sense-Knee-8, Transmitting Coil B: body, Repetition Time: 5000, Echo Train Length: 32, Flip Angle: 90, Acquisition Number: 8, Space between Slices: 5, Pixel Bandwidth: 1296, Matrix size: 256x256, Slice Thickness: 5, MR Acquisition Type: 2D, Software Versions(s): 2.6.3 & 188.8.131.52.
All 9 vials were taped together and scanned using Carr-Purcell-Meiboom-Gill (CPMG) sequence with and without Spectral Saturation Inverse Recovery (SPIR). Imagings of the water-oil cylindrical vials were repeated with the phantom in a horizontal position. The MR imaging studies were performed with a 3.0T MR imaging system (Achieva; Philips) operating with Release 2.6 software. The same scaling factors were applied to each image in the multi-echo acquisition. In order to investigate the influence of echo space on the T2 curves, the phantom experiments were independently carried out with a matrix of 256 x 256, TR of 5000msec., echo train length of 32, echo spaces of 10, 20, 30 and 40 millisecond respectively using standard CPMG and SPIR for fat and water. All scanning was performed under the supervision of a technologist (H.J., with over 15 years experience in MR imaging).
Image Processing and Analysis
The raw image data were exported from the imaging unit for off-line reconstruction and were converted to a format suitable for analysis by using software (MATHCAD, version 2000i, Parametric Technology Corporation (PTC Inc) Needham, MA). The determination of T2 is carried out with signal intensity values obtained using both standard CPMG and SPIR fat suppression sequences that utilizes the slice selective 180-degree pulses with full phase rewinding in between the refocusing pulses. The proton density and apparent T2 maps were estimated using pixel by pixel fitting curve of the signal intensities obtained with 32 echo times (TE) to a bi-exponential decay model. The increasingly T2-weighted images can be created by combining the 32 echoes of each slice into consecutive groups of four and calculating the geometric mean of each group. Algorithms provided by my adviser, the primary investigator for this project [H. J.] and implemented in MATHCAD software. The molecules containing fat and water will consist of oscillating signal intensity as a function of echo time. Simplified signal intensity formulas obtain from two components of water and fat by modeling water-olive oil mixture can be summarized as below:
SI = Signal Intensity, A = free protons, B = bound protons, W = water,
f = fat, N(H) = Proton Spin Density, TEi=Echo Time (i =1, 2 ...32),
Magnetization Transfer (MT): Selective presaturation RF can excite the tightly bound protons; the bound spins may transfer some of their magnetization to the free spins and thus reducing signal contribution from the bound pool. The reduction in signal will be proportional to the amount of magnetization transfer to the free spins in various molecular categories.
Statistical analysis was performed by using software (Microsoft Excel, version 2007). In paired two sample t-Test, Pearson correlation coefficient were used to examine the association between T2 values obtained from two similar situations that differ only in one condition. Differences with p-values less than 0.05 were considered statistically significant.
In this study for all given echo spaces, T2 values showed strong correlation with TE values even though that some correlation were not as strong as others but the range is generally very good (0.99-0.88, p-values < 0.05) for the standard CPMG sequences. T2 values obtained using SPIR fat suppression showed mixed result with strong correlations (significant p-values) with TE values at lower oil fractions, and differing only in some T2 values that were obtained with SPIR fat suppression at mid to high levels of oil fraction where the p-values that exceed the threshold 0.05 value. Both in SPIR fat suppression and standard CPMG, T2 correlation with TE diminished but remains relatively high as echo space (ES) is increased. T2 values are generally higher when there is smaller fraction of olive oil compared to water; the T2 values decreases as oil fraction increases but then stabilizes as the oil fraction increased, and however T2 cannot be said to correlate with the percentage of oil fraction. Statistically, echo spaces have some discernable effect on the T2 values which disappeared with increasing TE and T2-weighting.
In our study, shorter echo space produce compact sequence timing and had better resolution with reduced artifact; and in most clinical environment, information from T2-weighted images are view and analyzed in terms of qualitative nature by comparing observable differences in one or two images (see figure2). In this study, we were able to reconstruct T2 maps and measure the T2 values in each vial for both standard CPMG and SPIR fat suppression with similar parameters. The result showed large increases in standard deviations at increased echo spaces, which is an indication that there are large variations of T2 values obtained when using increased echo spacing at 3.0T. There are lower standard deviations at shorter echo spaces, thus less variation of T2 values. There are significant different between standard deviations from standard CPMG and standard deviations from SPIR FS, the deviations from standard CPMG are larger than those from SPIR FS (see figure7 A and B).
In a paired two sampled t-Test for T2 values obtained using ES of 13msec for both SPIR FS and standard CPMG, and using ES 10 msec (ETE are 65 and 50 msec. Respectively) for both SPIR FS and standard CPMG, their Pearson correlations at these given echo spaces are (0.97 and 0.84 respectively, with p-value < 0.05). This is to say that SPIR FS correlate well with Standard CPMG.
Zero signal intensities were recorded for pure olive oil in all the echo spaces using SPIR FS, which is good indication of the fat suppression. An effective fat suppression that can be translated into in vivo hepatic studies could provide a mechanism for quantifying T2 relaxation measurements. See figure1
Some of the effects seen on T2 can be attributable to magnetization transfer and others to the oil fraction effect. A two factor without replication ANOVA has p-values < 0.05, which is to say the length of the echo spaces has no experimental significant effect on both the standard CPMG and SPIR FS T2 values. See Table1 and Table2
In a separate paired sample t-Test for average T2 values using 3 short echo spaces (8.5, 10, & 13msec) for SPIR FS, their correlation range were 0.98 - 0.89 with p-values greater than 0.05. While a similar paired sample t-Test for T2 values for intermediate ES (20, 30, & 40msec), their Pearson correlation range were 0.95 - 0.73 with p-values < 0.05). There are observable stabilization on the values of T2 as the echo space is increased which is an indication of increased T2-weighting and presume less T1-w effect. But there is clearly some unexplained effects due to oil fraction and perhaps magnetization transfer; T2 values were higher for smaller fractions of olive oil and stabilizes as oil fraction increased. Thus a significant improvement can be achieved with increased ETL (shorter ES) in the T2 estimation. There are also increased observable noises as signal intensity diminishes with increasing T2-weighting (see figure3).
There is evidence of signal loss both in short echo space, as well as intermediate echo spaces for water suppression sequences. Fat has short T1 and T2 relaxation times and may experience signal intensity loss during water suppression due to water long relaxation times. In this study we are primarily interested in fat suppression rather than water suppression that can be translated to human studies of chronic liver diseases. In both fat suppression and water suppression, the fat signal diminishes as the TE is lengthened.
This study allows the simultaneous image acquisition of water-oil mixture as well their T2 maps. This study support T2 values stabilization and as such support the use of FSE imaging as one way to overcome T2* sharp-drop effects and enables relative assessment of oil and water signal components. It also support T2 measurements as a potential for providing information on the relative level of fat content in a water-oil mixture which could enable reliable assessment of hepatic tissue as a whole in a clinical study. But no clinical correlation to NALD and hepatic steatosis could be drawn from this study alone. This study support the use of shorter echo spacing to optimize multi-echo sequence, an optimized FSE is a way to accurately sample and estimate T2 values. Study support loss of signal strength at long TE especially for short T1 and T2 molecules such as fat; long ETE could lead to potential loss of signal intensity in lipid base molecules. Study support that T2 values from standard CPMG correlate with T2 values from SPIR FS and as such support the use of SPIR FS in a FSE for assessment of oil-water content. This study suggest in both cases of SPIR FS and standard CPMG that T2 values are not unique and as such represent some level of difficulty for efficient and accurate measurement of T2 values. And in both SPIR FS and standard CPMG, shorter ES can help minimize standard deviations in T2 measurements. Sequence optimization can mean different things to different people, from minimizing the scan time in order to increase throughput to maximizing the amount of information produced in order to increase pathological diagnosis. Research technologist and Radiologist must then strike balance between many competing objectives in order to reach the desired result. The expected increase in SNR from migration to 3.0T magnetic field strength and from combination of multi-echo sequence can be traded off to further reduce scan time with various acceleration techniques and reduce SAR that can be generated from high impact RF-pulses.
Given a fixed ETL, one will almost want to use as short an ES as possible in order to increase number of slices, improve the contrast control of the images and reduce image blurring. ES is directly related to the desired ETE, along with desired bandwidth. ES is a parameter that is fixed within a sequence; ETE on the other hand can be programmed through k-space reordering to achieve the desired result. K-space reordering is an administrator controlled function within most system's software and this can be reordered during each data acquisition. The k-space reordering is limited only by the system's hardware and software capabilities in placing data anywhere within the k-space and utilizing such data to reconstruct an appropriate image. ES is therefore an important parameter that could be employed to optimize TE during T2-contrast imaging. And given the importance of TE for quantification of T2 values, ES can be said to be an important factor in accurate calibration of T2 values, should quantification of T2 be deemed necessary for accurate estimation of the relative levels of fat fractions during many and particularly early stages of nonalcoholic liver diseases.
Our study is limited by how closely the use knee-coil receiver to model abdominal receiver coil. The cylindrical vials phantom was closely held together and placed in a cylindrical knee-coil receiver; abdominal receiver coils are surface based and may not match exactly with knee-coil. MRI is inherently three-dimensional and no relationship between imaging plane and the apparatus; however abdominal coil was not available for this study. Oil and water have been used in many research studies to model lipid-water study, we used olive oil which is known to have many generic versions and the quality is not always assured in many cases. Fat is known to have complicated chemical spectrum that contains a number of different spectral components, CH3, CH2, CH2COOR, and CH=CH groups which collectively represent the total fat signal . Even with a narrow band SPIR FS cannot completely saturate signal from fat without affecting signal from water. Any phase interaction between different fat components can add considerable complexity to the observed signal variation at a given TE. SPIR FS usually takes more time than gradient-echo based sequences; it requires high RF power thus increased SAR levels.
Given a long TR, shorter echo spacing is a method that can be used to optimize echo time and adequately sample T2 values. Short echo space can help shrink large standard deviations and large variations in T2 measurements. Spectral presaturation inverse recovery fat suppression (SPIR FS) T2 measurement correlate with Carr-Purcell-Meiboom-Gill (CPMG) T2 measurements and as SPIR FS can be used to estimate relative levels of lipid fractions in a water-fat study.